Always looking forward to returning to my home away from home, Silicon Valley, I jumped at the chance to attend my first ever hardware conference, AI and Edge AI Hardware 2024 put on by Kisaco Research (#AIHardware2024). The event took place September 9-12, 2024 in lovely San Jose, California, the heart of Silicon Valley. I was excited to see what innovations were happening on the AI hardware side. Boy, was I not disappointed! We mostly hear about all the fresh advances in GenAI model development, but now I see we’ll be hearing a lot more about the rapid evolution of AI-accelerated hardware to support these growing models at scale. AI hardware is evolving rapidly, and I’m not talking about NVIDIA Blackwell!
After picking up my badge, I checked the schedule of the sessions I was going to attend and then instinctively headed to the exhibit hall where all the companies were presenting fresh technologies. I knew that cheering on the vendors present would be the best employ of my time at the show. I learned SO MUCH after talking to most of the companies present.
Perhaps the most essential thing I’ve learned and heard over and over again is that it’s because GenAI training and inference have hit a wall. After all, how many GPUs can you buy? The good news is that I’ve talked to a number of companies that are thinking outside the box, developing alternatives to GPUs, each with a significant capability vector and talent density – think Brain, Furious, SambaNova, Grok, New Reality, Etchedplus many more. NVIDIA currently has a monopoly on training (90%) and inference (40%), so the industry is ripe for change. Many of the fresh hardware-accelerated AI solutions are tailored for inference—as the saying goes, . The biggest thing I took away from attending this conference is that change is in the wind. NVIDIA better start looking over its shoulder, because the company will soon have some earnest competition (not to mention the company’s antitrust concerns).
In my role as a member of the press, I try to make observations. One thing I noticed was the number of interviews taking place during the show. I saw many stand-up interviews taking place just outside the conference rooms. I also witnessed many strategic meetings between company executives, partners and investors. In the obscure corners of the hotel lobby bar, I overheard plans for fresh products, integrations, partnerships and investment rounds. It was refreshing to witness all the excitement, with memories of heady days past, and it led me to see how spirited the AI hardware space has become recently.
You can find a number of recorded presentations from speakers from the AI Hardware & Edge AI Summit conference HERE.
Location and place
Silicon Valley proved to be an appropriate location for the conference, as most of the companies offering fresh AI hardware solutions came from this region.
More than 1,300 participants were accommodated in one place, Hotel Signia by Hilton on Market Street. The entire second floor of the hotel’s conference room was dedicated to the show. It seemed like a space that would accommodate everyone involved. The hotel was also particularly tech-savvy – I didn’t have to interact with a single human being during my entire stay, as my reservation was managed 100% through the hotel app (I didn’t even need a room key, the app opened the door for me when I needed it!).
Exhibition hall
The exhibition hall was the focal point of the conference. It was routinely packed with companies in their booths along with attendees. There were also additional booths outside the hall in the adjacent corridors. Walking through the conference area, it always seemed to be crowded with people, which for me only added to the excitement.
Below is a tiny, selected list of companies from the conference I announced earlier.
The endless coffee and snacks in the exhibit hall were much appreciated. One of the few complaints I had about the event was the complete lack of seating to meet with companies, colleagues, fresh friends, or even have lunch. I noticed that many attendees had to sneak downstairs with their lunch/drink plates to the hotel lobby to find a seat.
Interviews
The main goal of my participation in the AI Hardware & Edge AI Summit was to conduct a series of video interviews with representatives of the following companies: Intel Labs, N-scale, MinIO, Hitachi Projects, EricssonAND Unilever. Each interview will appear here on insideAI News as an upcoming article. All of the interviewees were amazing and such good representatives of their companies. They all demonstrated a wealth of knowledge in their specialized topics. I learned a lot from interviewing them (and I hope you did too!).
The way home
I had a great time at the conference. I think the current focus on AI hardware was the biggest draw. I plan to do some earnest research on many of the companies I met with. I came away with a fresh and improved perspective on how the AI field is evolving.
As I waited for my return flight at San Jose Mineta International Airport (SJC), I found myself missing “The Valley” again until my next trip north (I live in Los Angeles). I am never disappointed with what I take with me when I visit this unique place on Earth, where most AI comes from.
I highly recommend this event to anyone in the AI space or looking to get involved. Getting a hardware perspective is essential now that hardware is frosty again! I think I’ll definitely add this event to my annual conference list!